Technology Update: RapidKL to provide real-time info on services

TRANSIT took note of this article in the Malay Mail back in February, which describes technology and service improvements by RapidKL which will see realtime service info brought to screens & online.

This form of service tracking, using GPS-based technology, is a welcome improvement because it allows public transport users to understand services in real-time and make transport decisions based on that information.

The reason for the nearly 2 month gap between the article & the post is that it has taken a great deal of time to analyze the articles and look for resources to help explain fleet management and GPS tracking.

We warn you now that this will be a very long, detailed and technical post – but it is definitely worth reading. So if you have any questions or have trouble understanding any details, please let us know or get clarification.

RapidKL to provide real-time info akin to train stations (Malay Mail)
Shahrim Tamrin
Friday, February 11th, 2011

KUALA LUMPUR: Bus commuters can soon look forward to getting real-time information on how long it would be before the next bus arrives in a system similar to the electronic schedule display boards at LRT train stations.

[TRANSIT: We need those information display boards at the entrances to the LRT stations as well so passengers can know if their trains are on schedule, how much time they have, if there is a delay, etc.]

Initial trials have begun and Syarikat Prasarana Negara Bhd group managing director Shahril Mokhtar is hopeful the information displays could be active at all designated bus stops before year-end.

He said the Fleet Tracking System (FTS) for the RapidKL bus fleet in the Klang Valley would use Global Positioning System (GPS) devices installed in the buses.

The Malay Mail understands the GPS — fitted to support the bus estimate time of arrival (ETA) public information system — has been enforced in Penang by RapidPenang, another subsidiary of Prasarana.

It had been installed gradually in Rapid Penang’s 300 buses for the past three years.

Bus public information system

Its main LCD board displaying ETA information of RapidPenang buses at Komtar building was constructed last May and received good feedback from bus commuters on the island.

“From June, our buses will be fitted with GPS devices. They will be linked to the Bus Control Centre (BCC) and an estimate of how long the next bus would take to arrive would be relayed to LCD screens at key bus stations like Pasar Seni, KLCC, Bukit Bintang, Jalan Tunku Abdul Rahman, Lebuh Pasar, Jalan Dang Wangi, Jalan Silang and other main locations around Klang Valley,” said Shahril.

“The such real-time bus arrival estimates would also be displayed at LRT stations so disembarking passengers know how long they have to wait for the next feeder bus.”

[TRANSIT: The information should also be placed online on the rapidkl website. This can also be arranged through a service like Nextbus or Google Transit.]

He continued: “One major grouse commuters have with RapidKL is that buses were not prompt due to traffic congestions and such. Not knowing when buses would arrive causes commuters to waste time.

“With this system, commuters can decide how to better manage the waiting time. It also paves the way for more informed travel decisions.”

[TRANSIT: No disagreement there. However, it must be pointed out that having the information about bus arrival is only part of what is needed to improve public transport service. Fleet management is another important factor.

Knowing when the next bus is to arrive only helps passengers make better transport decisions if passengers have multiple options available, or have the time & interest to wait.

That said, knowing is better than not knowing – which is the current situation.]

Shahril said FTS for RapidKL is a catalyst to support Prasarana’s on-going plans to transform public transportation in the city.

“We are also going to introduce short-messaging (SMS) and website services for the public to estimate when the next bus arrives at a specific stop for any designated service.”

The SMS and website services concept was similar to the LCD displays as it too will be integrated with the FTS comprehensive data at BCC.

On fleet expansion, Shahril said Prasarana would purchase 470 new buses in stages by September to complement the existing LRT feeder buses and RapidKL fleet.

“Some 30 per cent of the 470 buses will beef up the feeder bus service for LRT stations and the rest would increase the current 1,100 buses serving the Klang Valley.

“With FTS and the public information system, we are seriously focusing on improving our on-time performance, service frequency and customer satisfaction. Again, this is all part of Prasarana’s commitment to increase the public transport ridership.”

TRANSIT Says:

When we first read about the Fleet-Tracking system we were quite pleased to hear about the investment in information and communications technology – but at the same time, we were concerned that this technology by itself will not resolve many of the issues that exist with bus services – namely scheduling, fleet management, and buses getting caught up in traffic jams.

If these issues are not resolved, we would end up with a situation similar to the ITIS message boards – technology that can only inform you about traffic congestion but cannot do much to reduce congestion.

Strangely, enough, our comments and questions about fleet management and traffic issues & challenges were not well-received by some.

In any case, the next section gives you an idea of how the public information portion of the Fleet Tracking System can be useful.

Overview of the public information system

A SIMILAR public information system for estimated arrival time for buses is already available in Singapore, India, South Korea, Hong Kong as well as major cities in Europe, US, Canada and South America.

The system is known as the Intelligent Transportation Systems Programme or Advanced Public Transport System. GPS-enabled devices allow real-time tracking of bus speed along the designated route so that updated data can be fed to passengers waiting at bus stops ahead.

Interactive signboards display next-bus arrival time information, which can also be sent via text message and smartphone applications — thus expanding public transport options for commuters.

[TRANSIT: Of course, the depends on the availability of options]

Such information also helps improve efficiency of bus operations as the ETA data will also be displayed in buses for passengers wanting to change to a different bus at an interchange.

TRANSIT Says:

This information from Nextbus gives you a good visual idea of how their system works.

Now, one of the major issues in the Klang Valley is the lack of integrated interchanges that will bring together multiple public transport services in a concise, complete and effective manner.

For example, a public transport user in SS16 Subang Jaya who wants to go to KL has the following options available to him or her:

• KTM Komuter from Subang Jaya KTM station;
• RapidKL bus U67 to Pasaramakota (Klang Bus Stand via Fed. Hwy);
• RapidKL bus U76 to KL Sentral (via Old Klang Road);
• RapidKL bus U623 to Kelana Jaya LRT station, then the LRT to KL;
• Taxi to Kelana Jaya LRT station, then the LRT to KL;
• Taxi to KL directly.

Now, all of those options are viable, but their use depends on many factors. One factor is availability. The KTM Komuter service is the most “available” and the easiest to access, but the commuter needs to know if the service is on time. Another issue with the Komuter is the overcrowding and cancellation of train services.

The RapidKL LRT from Kelana Jaya is another option and the LRT is more reliable than the KTM Komuter service, and faster, but it costs more for the trip and the feeder bus trip. Also, the bus stop for U623 (the feeder bus) is located across busy Jalan Kemajuan Subang.

Another factor is timeliness. The bus options (U67 & U76) are good, direct options but require waiting for the bus to arrive – and not knowing when it will arrive. Not to mention, buses are much slower because of the large numbers of bus stops as RapidKL does not offer express options on most of the corridors yet.

Then there is a cost – taxis are a fast, direct option (either to the LRT station or to KL directly) but they are costly – assuming that you can get the taxi driver to agree to take you to your chosen destination.

Now, the information system would at least tell you if the best choice would be to bear with KTM, or wait for the feeder bus to the LRT station, or just take a taxi – but if your U623 bus is stuck in a traffic jam on the LDP at Sunway, then there isn’t much you can do except wait.

The other big question is how would you keep updated when you are mobile? We are told that there would be message boards at various bus hubs, but let’s be honest – we have seen many RapidKL buses with LED destination signs that are not operating – and those are signs that are under the control of RapidKL 100% of the time. If they cannot keep those working ….

The other option is mobile technology – your web-enabled or app-enabled* smartphone or your handphone for SMS alerts.

*Many application developers are already taking advantage of the open data for development of their own apps at the service end. TransSee is but one example, and there are google groups of public transport-related app developers.

However, this assumes that RapidKL and RapidPenang will make the data open and available to the public.

Portland is one of many cities that have joined up with the open data concept and are making information available to the public. The video below gives an idea of how things work.

The next section discusses some of the specific benefits of the Fleet-Tracking System including those for the operator.

Benefits

• Real-time information to commuters through LCD boards at bus-stops, internet and SMS;
• Enhancing commuter satisfaction;
• Better bus scheduling;
• Reduced out-of-route miles;
• Basic communication between driver and control room in emergencies;
• Quick replacement in case of breakdown/accident en-route;
• Effective control over the drivers and checking on skipping of bus-stops;
• A check on over-speeding;
• Increased punctuality;
• Automation of fleet operations minimises human intervention;
• Improved fleet utilisation, leading to better services

TRANSIT Says:

When we first heard that RapidKL was introducing a fleet tracking system using GPS information, we immediately wondered how the FTS would benefit RapidKL in terms of route planning, service planning and fleet management – because let’s face it – you can have all the technology & hardware that you can think of but if you are not managing the information & using it for the benefit of customers (as well as keeping service available and reducing waste) then the investment was pointless.

A good resource on the issue of service cost & quality can be found at the blog of Steve Munro, especially at this link here.

TRANSIT has become familiar with some of the fleet management technology & use of data that is available through GPS tracking (and open data) to help understand bus & rail service tracking.

Fleet Tracking:

Fleet tracking is an analysis of the actual frequency of vehicles based on GPS (or other forms of data). Steve Munro explains the chart showing service on the St. Clair LRT line in Toronto, on 5 July 2010:

This format of chart dates back to the late 1800s when it was invented in France for building railway schedules, and it is still used today for that purpose.

Time flows from left to right beginning at 4:00 am and running over 8 pages to 4:00 am the following day.
Distance flows on the vertical axis with St. Clair Station (the eastern terminal) at the bottom and Gunn’s Loop (the western terminal) at the top. Major intersections are noted by lines across the chart.

All positioning data are GPS based reported every 20 seconds from the vehicles. Wiggles in the lines show the variation in speed, and horizontal stretches indicate that a vehicle is stopped. This is notable at the termini and at St. Clair West Station, but also for shorter periods at intersections where it is possible to distinguish the nearside and farside stops. (Occasionally, a vehicle reports a rogue location from its GPS. These are usually trapped out in the data mapping process because they are not reasonable locations for a vehicle on the route. Sometimes, an erroneous location will appear that is near enough to the route to appear as “real” and this can cause a larger wobble in the chart line for the affected vehicle.)

Basically, what you should see in the fleet movement chart is a set of relatively consistent, evenly-spaced lines that follow the same pattern. Any change in the pattern (horizontal line, wider spacing, narrower spacing, etc. shows that something is happening in real-time to the actual service.

For example, a route that is consistently showing a horizontal line (see page 3 of the chart) might have a very slow driver or driver who is not operating his vehicle (on break or just not working) – or there just might be too much traffic causing congestion.

In another example, 3 lines that start to come very close together might indicate that 3 vehicles are bunching – meaning that there would be a longer wait for passengers at either end of the bunch.

In the control centre, the controllers can take action – ask a driver to drive faster or slower (or increase or decrease train speed in the case of the Kelana Jaya line) – as well as collect data that could justify changes that need to be made.

Link Times:

This next set of chart analyzes “Link Times” which effectively refer to how quickly a vehicle is able to go through a segment of the route, getting from point to point.

This chart shows ‘link times’ for westbound vehicles on the St. Clair route on 5 July 2010. A full explanation and comparison of Link Times on the St. Clair route in 2 different months can be found here.

When looking at charts showing “Link Times” the vertical axis represents the time between the two points mentioned in the title, while the horizontal axis looks at the period being analyzed (a particular hour, day, week, month).

You should look for a consistent pattern of connections that are approximately the same height. This means that the vehicles are relatively close together. Small variations in height represent things like traffic lights, small congestion, a car parking, waiting for a passenger etc.). Large variations represent significant delaying factors like double parking, traffic congestion, emergencies, mechanical failure, etc.

For example on page 3 of the chart above at 15:00 you can see that trip times for 3 vehicles have increased from 3.5 minutes to over 6 minutes – suggesting a delay that caused a backup of 3-4 vehicles. On page 7 you can see that one trip (at approximately 13:00) took more than 9 minutes, well over the average of 1.5 minutes – but the next trip was back at the average. Maybe the earlier driver stopped to get a coffee.

Also on page 7 you can see a period of significant delay between 10:30 and 11:00 where no streetcars were running – represented by the shallow line – but after 11:00 trip times improved and moved back to the average.

Headway Analysis:

This next set of charts allows the comparison of ‘real’ headways (time between vehicles – or frequency of service) against ‘scheduled’ headways on the St. Clair LRT. More information on Headways can be found here.

According to Steve Munro, the scheduled headways on St. Clair are:

• AM Peak: 2’55″
• Daytime: 6’00″
• PM Peak: 3’30″
• Early Evening: 6’00″
• Late Evening: 9’00″

What is very interesting is that despite the data above (which shows no major delays on the St. Clair route with a few minor exceptions, the variations in vehicle headways (or service frequency) on 5 July 2010 is quite significant – both in comparison to their schedules and to each other!

Take a look at the same westbound data for 5 July 2010 on the chart here.

What jumps out at you on the first page is a vehicle that sits for more than 24 minutes – but it is quite possible that there is a mechanical issue and this is a terminal station. So if we go to page 3 we can see major headway gaps appearing right away and staying present through different periods of the day.

What is especially interesting is that the largest headway gaps (the most vertical of the jumps) appear outside of the peak period times – especially in the later evening hours. For example, on page 6 there are significant variations after 09:30 and before 13:00. During the period from 15:00 to 18:00 the average headway decreases but if you look closely you can see that every other vehicle has a very large headway – a sure sign of vehicles bunching & delaying others.

On page 7 you can see significant gaps start to appear after 18:00 – this area is closer to the end of the route so operators may decide to drive slower (because their scheduled rest time might allow them a window of time to get to the terminal) or faster (in order to arrive earlier so they have a longer rest time at the terminal).

Conclusion

We have only been able to give you a very microscopic sample of what kind of data can be available to fleet managers and transport planners using existing fleet-tracking technology.

And ideally, this post should give you an idea of how complicated fleet management is – especially in an urban public transport situation where there are many, many variables.

TRANSIT would be highly interested to see data from RapidPenang (which has a GPS tracking system) showing their headways & link times – and working to find areas where buses are going ‘off track’, driving too fast or too slowly, getting caught by traffic congestion and traffic lights etc. And similarly, we would be interested in seeing test data from RapidKL’s GPS testing.

In all cases, these data can be used to argue for better funding, more training, more infrastructure (bus lanes for example – although they are not a panacea for congestion problems).

Most importantly, the use of GPS & fleet tracking technology allows bus operators to make their service more efficient and improve the customer experience.

And that is something we can all agree on.